Mapillary交通标志数据集（国外）

Mapillary公开数据集包含来自世界各地的100,000张街景图像，涵盖300多个交通标志类别，共标注超过320,000个交通标志实例。数据集在天气、光照、摄像头传感器和视角等方面具有高度多样性，适用于训练和评估交通标志检测与识别模型。该数据集广泛应用于自动驾驶领域的交通标识识别任务，助力提升复杂道路环境下的感知能力。The Mapillary Public Dataset contains 100,000 street view images from across the globe, covering over 300 traffic sign categories with more than 320,000 annotated traffic sign instances. The dataset exhibits high diversity in terms of weather, lighting conditions, camera sensors and viewing angles, making it suitable for training and evaluating traffic sign detection and recognition models. It is widely applied to traffic sign recognition tasks in the field of autonomous driving, aiding in enhancing perception capabilities in complex road environments.

搜集汇总

背景与挑战

背景概述

Mapillary交通标志数据集是一个大规模、多样化的公开数据集，包含来自世界各地的100,000张街景图像，标注了300多个交通标志类别和超过320,000个实例。该数据集在天气、光照、摄像头传感器和视角等方面具有高度多样性，专门用于训练和评估交通标志检测与识别模型。